AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Reasoning-Guided Part-Level Visual Grounding via Reinforcement Learning

Kazi Sajeed Mehrab, Hani Alomari, Najibul Sarker, Zaber Ibn Abdul Hakim, Chia-Wei Tang, Anuj Karpatne, Chris Thomas

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Chain-of-thought multimodal reasoning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train with part-aware GRPO and OP-HRG prompt instead of zero-shot prompting (ablation)OP-HRG · InstructPart (test)gIoU: 31.45 -> 75.56Improved
add hierarchical grounding (object-first, containment, object hint reward) (ablation)OP-HRG · InstructPart (test)gIoU: 73.77 -> 75.56Improved
add reflective refinement (self-critique and improvement reward) (ablation)OP-HRG · InstructPart (test)gIoU: 73.98 -> 75.56Improved
add OP-HRG structure and part-aware rewards over plain localization rewards (ablation)OP-HRG · InstructPart (test)gIoU: 70.32 -> 75.56Improved
adapt VisionReasoner with GRPO on part data only (ablation)VisionReasoner · InstructPart (test)gIoU: 59.38 -> 62.33Improved
swap SAM2 mask decoder for SAM3 (ablation)OP-HRG · part benchmarksgIoU: 75.56 -> 75.77 (InstructPart)No clear effect
add active visual perception at 4B (ablation)OP-HRG · part benchmarksgIoU: PartImageNet 56.87 -> 55.98; InstructPart 75.56 -> 76.34Mixed
scale backbone from 4B to 8B with AP (ablation)OP-HRG · PascalPartgIoU: 39.04 -> 40.28Improved
evaluate with plain single-answer prompt instead of reflective prompt (ablation)OP-HRG · InstructPart (test)gIoU: 75.56 -> 75.40No clear effect
part-centric RL trainingQwen3-VL-Instruct-4B · RefCOCO/+/g averageaccuracy (%): 89.7 -> 86.5Hurt
OP-HRG-trained 4B vs VisionReasoner 7BOP-HRG · ReasonSeggIoU: 63.6 -> 69.6Improved
remove external baseline from improvement reward (ablation)OP-HRG · training dynamicsinitial-answer IoUHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy